Methods: A survey was collected from a random stratified sample of incarcerated women (N=574) in one Midwestern prison. Completion of the survey was voluntary and anonymous. This analysis focuses on variables related to current /past criminal behavior; victimization and perpetration of physical abuse and intimidation by/toward partners and non-partners; and personality inventories (e.g., anger, conduct problems). Hierarchical and k-means cluster analyses were used to group women based on patterns of perpetration toward partners and non-partners. Multinomial logistic regressions were conducted using victimization histories as predictors for perpetration cluster membership.
Results: Women were grouped into one of five perpetration clusters: 1) no perpetration, 2) physical abuse and intimidation toward partners, 3) physical abuse toward non-partners, 4) intimidation toward non-partners, and 5) high physical abuse and intimidation toward partners and non-partners (n=291, 109, 57, 60, 57 respectively). Clusters showed significant differences in substance use, anger and personality. For example, women in cluster #1 and #2 (no perpetration and partner perpetration) had significantly fewer substance use disorders than women in the high perpetration group (cluster #5). Similarly, women in the no perpetration group (#1) had significantly lower scores on conduct problems, antisocial personality behaviors, and instrumental /expressive anger than women in all other clusters. Victimization histories provided additional context and predictive power for understanding women's perpetration. Not only did victimization predict cluster membership, but also we found that the specific type of victimization (i.e., physical abuse or intimidation) and assailant (i.e., partner or non-partner) mattered. For example, women who experienced physical abuse by non-partners were more likely to perpetrate physical abuse toward non-partners (cluster #3; OR=14, CI=3.6-52.3). Several significant interactions emerged, providing a complex picture of how cumulative victimization history predicts perpetration cluster membership.
Conclusions and Implications: Although women's engagement in violent behavior, particularly violence that is criminal, is far less than men's, it is important to understand the patterns and likely motivations of their aggression within and outside of partner relationships. Social work practitioners and researchers interested in reducing and/or preventing sequalae associated with violence against women must consider both women's perpetration and victimization in the development and implementation of interventions designed specifically to address the cumulative impact of trauma on women's lives.